Neuronal morphological plasticity: mechanisms and meaning / Plasticité morphologique neurorale : mécanismes et fonction
Séminaire FSER organisé par Bernardo Sabatini (Harvard Medical School) et Tobias Bonhoeffer (Max Planck Institut) du 21 au 26 mai 2007
In this meeting, were discussed the mechanisms that underlie activity-dependent alterations of neuronal structure and how changes in neuronal structure might impact the ability of the cell to receive, process, and transmit information.
Veronica Alvarez, Tobias Bonhoeffer, Frances S. Chance, Graeme Davis, Michael Frotscher, Kristen M. Harris, Michael Häusser, Christiaan Levelt, Christian Lohmann, Christophe Mulle, Dominique Muller, Elly Nedivi, Thomas G. Oertner, Bernardo Sabatini, Antoine Triller, Noam Ziv
Animals store information and learn new behaviors through manipulating the connections between neurons in the brain. The regulation of the function of synapses is typically also accompanied by structural changes. In this meeting, we discussed the mechanisms that underlie activity-dependent alterations of neuronal structure. In addition, we discussed how changes in neuronal structure might impact the ability of the cell to receive, process, and transmit information.
Dr. Sabatini presented an analysis of the control of PSD95 trafficking in the synapse and its role in spine morphological plasticity. Stability of PSD95 at the synapse was regulated by its first two PDZ domains and by phosphorylation by CAMKII at serine position 73. Phosphorylation of serine 73 also prevented plasticity associated spine growth. Lastly, analysis of spine head calcium transients was used to argue that the spine neck poses a significant barrier to current flow and thus permits large, compartmentalized electrical signals within the spine head.
Dr. Frotscher presented an electron-microscopic (EM) analysis of synapses in hippocampal organotypic slices that had been process via a rapid, high-pressure, freeze and cryosubstitution. He analyzed the spine apparatus (SA), a membranous structure present in some dendritic spines of hippocampal pyramidal neurons. He found that synaptopodin knock-out mice lack the SA and show alterations of long-term potentiation (LTP) of the CA3 to CA1 synapse. In addition, the mice have decreased performance in hippocampus dependent tasks. At a functional level, spines of synaptopodin knock-out mice have slowed clearance of action potential evoked Ca transients.
Dr. Triller examined the movement of neurotransmitter receptors on the surface of neurons. Individual receptors were labeled by quantum dots via an antibody and tracked by fluorescence microscopy. Glycine receptors were following in dissociated neuronal cultures and were found to be diffusionally restricted via interactions with gephyrin. In addition, the movement of glycine receptors was rapidly regulated by changes in the activity of glutamatergic synapses. Lastly, Dr. Triller also presented data using high-pressure cryo-EM processing. Using tomographic EM, the presynaptic density was shown to contain inhomogeneous patches of active zone proteins as well as long tethers that linked distant synaptic vesicles to the active zone.
Dr. Lohmann presented data on the activity- and neurotrophin-dependent plasticity of hippocampal organotypic slice cultures. Individual neurons were loaded with Ca indicators through single cell electroporation. Application of BDNF triggers Ca transients within dendritic spines whereas inhibitors of BDNF signaling reduced them. Ca transients are also seen in thin filopodia, which grow in regions of low Ca signaling but are stabilized following increases in Ca.
Dr. Mulle analyzed the role of Kainate-type glutamate receptors (KAR) in synaptic plasticity. He demonstrated that the KA1 and KA2 subunits can form functional receptors only when associated with GluR5-7 subunits. At the mossy fiber to CA3 synapse, a portion of the synaptic current is carried by KAR. These currents have slow kinetics and can summate substantially during repetitive stimulation. Surprisingly, small levels of KAR agonists led to alterations in the slow after-hyperpolarization, possibly due to activation of a G-protein mediated cascade. The hypothesis is presented that the GluR6 subunit opens the ion channel whereas KA2 activates a G-protein. In addition, data was presented that presynaptic KAR modulate neurotransmitter release and plasticity induction.
Dr. Oertner presented data that the spine head operates as an isolated electrical signaling compartment that, during synaptic activity, is depolarized to near 0 mV. He demonstrated that blockade of AMPARs has a large impact on synaptically-evoked Ca influx, consistent with a large depolarization in control conditions that relieves Mg block of NMDARs. In addition, Ca influx through voltage-sensitive Ca channels seems to contribute to or regulate the synaptic signals. A model was presented that accounts for the data.
Dr. Davis presented an examination of the processes of synapse maintenance and disassembly at the fly neuromuscular junction (NMJ). His laboratory identified dynactin and spectrin in a screen for molecules that affected synapse stability. Mutations in dynactin greatly increased the percentage of retracted synapses. Mutations have recently been reported in this transport molecule in human ALS. Mutations in spectrin also altered glutamatergic transmission at this synapse.
Dr. Harris used EM to study spine and synapse structure in hippocampal slices that had experienced synaptic plasticity. The number of mushroom spines and filopodia increased post LTP induction. A model was presented in which filopodia are transformed into thin spine and then mushroom spines after LTP. Furthermore, the numbers of polyribosomes, sites of active protein translation, were increased in spines at 5 and 30 minutes after LTP induction but not at 2 hours.
Dr. Levelt examined the role of TrkB signaling in visual cortex plasticity, using an ocular dominance plasticity model. Intrinsic signal imaging was used to monitor activity in the cortex and measure spatial acuity to contrast inverting gradients. In addition, individual neurons were labeled with GFP and morphology of control versus TrkB-lacking neurons was compared. TrkB-lacking cells had decreased mushroom and stubby spines as well as increased thin spines and filopodia. There was a concomitant loss of synapses; however, this did not mirror the total change in spine density and was more closely predicted by the loss of mushroom spines.
Dr. Nedivi examined the plasticity of interneuron dendrites in vivo. In adult animals, the tips of interneuron dendrites showed growth and retraction over a period of days and weeks. Layer 1 neurons appeared more stable than neurons in lower layers. Preliminary data suggested that dendrite plasticity was increased by monocular deprivation. In addition, an analysis of the CPG15 knock mouse was presented. This animal showed widespread changes in synaptic transmission or structure in many brain areas.
Dr. Ziv presented data on the role of network activity in modulating synaptic remodeling. He used a long-term culture of very high density dissociated neurons that were continuously imaged using a confocal microscope. Activity was monitored using a multielectrode array and synapses were visualized due by the fluorescence of a PSD95-GFP fusion protein. Activity was manipulated by potentiating GABA transmission with a benzodiazepine. This triggered a transient drop in network activity that recovered over 10 hours. During this period PSD95-GFP puncta increased ~10% in size.
Dr. Bonhoeffer presented data concerning the morphological correlates of synaptic plasticity in a variety of systems. In organotypic slice cultures, patterns of activity that lead to LTP or LTD also lead to the growth or retraction of dendritic spines, respectively. EM analysis of new spines showed that these do not form an ultrastructurally mature synapse until 8-16 hours after appearing. In vivo, manipulations of visual activity by induction of a retinal scotoma triggered large increases in the turnover rate of spines. This effect was limited to the spines of neurons within the cortical area whose retinotopy corresponded to the retinal legion and disappeared within 2 months of the lesion.
Dr. Alvarez presented an analysis of the roles of NMDARs in regulating spine density and plasticity. She found that pharmacological blockade of the receptors increased the fraction of transient spines but had no effect on overall spine density. In contrast, physical loss of the receptor by RNA-interference led to increase spine morphological dynamics and a profound loss of spines over a period of 1-2 weeks. Spine density could be rescued by combined expression of NR1-1 and NR1-2 splice isoforms. She proposed that loss of NMDARs from the synapse is the final trigger for the retraction of the spine and synapse.
Dr. Chance presented a computational method to analyze the effect of one synaptic input on the spiking pattern of a neuron. She proposed that ‘receiver operator curve (ROC) analysis’ provided a method to characterize these effects in the face of changing spontaneous and basal firing rates. Furthermore, she presented a theoretical analysis of how the ROC curve is affected by shunting conductances and changes in synaptic parameters.
Dr. Hausser presented data linking dendritic morphology and synaptic plasticity. He showed the results of computational studies demonstrating that the dendrite patterning influences the propagation of electrical signals. Furthermore, he showed experimental data that the location of a synaptic input onto a dendritic tree determines the sign of plasticity that is induced by pairing with postsynaptic activity.
Dr. Muller presented an analysis of structural synaptic remodeling associated with plasticity. Time lapse analysis of spines in organotypic hippocampal slice cultures was used to quantify the rate of spine turnover and to measure the morphology of spines. He demonstrated that filopodia are highly unstable and unlikely to last more than 24 hours and that only half of new mushroom spines are likely to last more than 24 hours. Furthermore new spines are unlikely to have PSDs or express PSD95-dsRED until 5-19 hours after initial growth.
In conclusion, many links were found between the structure and function of neurons. Thus it was clear that the movement of proteins into and out of the synapse is limited by the structure of the dendrite and spine. Furthermore, the structure of the synapse and the neuron responds directly to changes in activity and synapses can be grown, reshaped, or retracted rapidly. However, surprises came from descriptions of exceptions to the general structure/function relationship. For example, most of us expected that the growth of a new spine would correlate exactly with the establishment of a new synapse. Beautiful evidence showed that this is not the case and that new spines do not gain a functional synapse for many hours after their initial growth. Thus it is clear to us that although the mechanisms that mediate cell growth in response to activity are being elucidated, we still do fully understand the functional meaning of the morphological changes.